The objective of this project was to evaluate the air quality aspects of prescribed burning by measuring fuel loading, fuel consumption, and the emission factors for gases, metals, and particulate matter (PM) on Department of Defense (DoD) facilities in the West and Southwest. These measurements provided input for a detailed model-based assessment of the regional air quality impact of prescribed burning. The measurements and model results provide DoD land managers and regional air quality agencies with the data needed to more accurately predict the air quality impacts of prescribed burning.

Technical Approach

The complex science issues associated with fires and their impact on local and regional air quality were handled by assembling a team of U.S. Forest Service and academic experts for fire spread, emission measurements, and emission modeling with personnel at DoD sites. The SERDP process added a scientific advisory committee to provide guidance on advancing the science and a focus on help needed by DoD resource managers. The technical approach involved both lab and field measurements.

Key measurements were made at the U.S. Forest Service lab in Montana where individual wildland fuels were burned during a total of 77 runs, including triplicate tests for many fuels. Forty-nine of the runs were southwestern fuel types. Gas phase emissions were measured with traditional methods and a new infrared tool, specifically developed for this project, which allowed measurement of many oxygenated hydrocarbons and ozone initiators. Instruments from the University of California, Riverside (UCR) particle science labs were added, and these enabled continuous emission measurements of the chemical and physical properties of the PM over the course of a fire: from the flaming to the smoldering phase.

Prescribed burns and emission measurements in the field followed to validate lab-measured emissions factors. The field equipment included an instrument-packed airplane that followed the airborne solid and liquid particulates and gases released from the fire for hours to learn what happens to aged smoke. Later a few exploratory trials in UCR’s atmospheric reactors aged smoke to learn if results would be similar to the aircraft findings.

In the final task, EPA air quality models were used to forecast the concentration and elevation history of gases and smoke released from a fire. These results were compared with SMARTFIRE and BlueSky, which are simpler models to run.


Improved fuel characterization and consumption data. Data for southwestern fuels such as chaparral and the Emory oak (Quercus emoryi) woodlands were scarce. Wildland fuels at three military facilities in California and Arizona were sampled, characterized, and compared with other work. Data included measured field and chemical properties as well as the consumption rates during prescribed burns. Collectively, these data provide improved values for fuel consumption rates during fire and are available for managers of military facilities and neighboring lands.

Improved emission factors and new test methods were the main emphasis for the project. Emissions data for gases, metals, and PM were generally lacking for the fuels burned in this project. Using EPA’s AP-42, Compilation of Air Pollutant Emission Factors, as a guide, the project produced new emission factors for criteria pollutants (CPs) and selected hazardous air pollutants (HAPs). The CPs included CO, NOx, SOx, PM2.5 mass and lead (Pb). HAPs included: aldehydes, ketones, ammonia (NH3), benzene/toluene/xylene (BTEX), and polycyclic aromatic hydrocarbons (PAHs). Formaldehyde was the main aldehyde measured in the fire. In addition, measurements made by SERDP project RC-1649 improved emission factors for previously undocumented oxygenated gaseous compounds and HONO, an important ozone precursor.

The RC-1648 team used real-time instruments to characterize the chemical and physical nature of the aerosols and PM. These new methods followed the path of burning from flaming to smoldering and reported the instantaneous change in particle diameter distribution, number, and aerosol composition. For example, the Aerosol Mass Spectrometer monitored the elemental chemistry and levoglucosan, a known marker for biomass fires. In addition to PM mass, the project team determined the elemental and organic fractions of the mass and 38 elements on the filter. Ions, including SO4, NO3, Cl, Br, Na, NH4, K and Ca, were analyzed on the filter.

One limitation of lab fires is that the combustion process is over in minutes and the products are quickly vented. In real fires, reactions take place over hours as the plume migrates downwind away from the source. In this project, an instrumented aircraft flew in the downwind plume to sample smoke that was hours old to learn about the transformations. However, aircraft monitoring is expensive, so RC-1648 carried out exploratory runs in an atmospheric reactor filled with combustion products and monitored the changes to the emissions under various conditions. The exploratory trials showed results similar to aircraft data suggesting further atmospheric reactor studies.

A significant advancement in data analysis of emissions from wildland fires came during the data analysis task. Traditionally data are fitted to combustion efficiency. RC-1648 showed that the percentage of total filterable carbon that is graphic in nature enables one to model the release of black carbon, brown carbon, and lighter molecules, like levoglucosan. This parameter is a surrogate for fire intensity and provided a surprisingly good fit.

Air Quality Modeling. This work involved modeling the emissions with the Community Multi-scale Air Quality (CMAQ) Model and forecasting the impact on air quality around Vandenberg AFB. Winter sun intensity limits ozone formation so gaseous values were low; however, the comparison of model and field data looked promising. Trials with BlueSky were better than SMARTFIRE and may offer advantages over CMAQ for land managers.


  • Developed improved models relating to fuel types, fuel loadings, and consumption estimates for Southwest chaparral and oak fuels.
  • Improved emission factors for CPs, select HAPs, and elemental factors in EPA’s AP-42 format for wildfire for DoD facilities with southeastern and southwestern fuels.
  • Field data showed DoD facilities should consider entrainment of dust into the plume before prescribed burns as significant amounts of lead and antimony were detected in the ground-based filter samples at Vandenberg AFB. Filters from laboratory burning of the fuels did not include these elements in any significant quantity suggesting a non-fuel source.
  • A new mathematical model was developed to predict the release of black carbon, brown carbon, and lighter molecules, like levoglucosan, from wildland fires.
  • Air quality modeling suggests that the BlueSky framework might be more useful to local land managers than the complex CMAQ.